1Git 2.8 Release Notes 2===================== 3 4Updates since v2.7 5------------------ 6 7UI, Workflows & Features 8 9 * "branch --delete" has "branch -d" but "push --delete" does not. 10 (merge 38a2559 ps/push-delete-option later to maint). 11 12 * "git blame" learned to produce the progress eye-candy when it takes 13 too much time before emitting the first line of the result. 14 15 * "git grep" can now be configured (or told from the command line) 16 how many threads to use when searching in the working tree files. 17 18 * Some "git notes" operations, e.g. "git log --notes=<note>", should 19 be able to read notes from any tree-ish that is shaped like a notes 20 tree, but the notes infrastructure required that the argument must 21 be a ref under refs/notes/. Loosen it to require a valid ref only 22 when the operation would update the notes (in which case we must 23 have a place to store the updated notes tree, iow, a ref). 24 25 * "git grep" by default does not fall back to its "--no-index" 26 behaviour outside a directory under Git's control (otherwise the 27 user may by mistake end up running a huge recursive search); with a 28 new configuration (set in $HOME/.gitconfig--by definition this 29 cannot be set in the config file per project), this safety can be 30 disabled. 31 32 * "git pull --rebase" has been extended to allow invoking 33 "rebase -i". 34 35 * "git p4" learned to cope with the type of a file getting changed. 36 37 * "git format-patch" learned to notice format.outputDirectory 38 configuration variable. This allows "-o <dir>" option to be 39 omitted on the command line if you always use the same directory in 40 your workflow. 41 42 * "interpret-trailers" has been taught to optionally update a file in 43 place, instead of always writing the result to the standard output. 44 45 46Performance, Internal Implementation, Development Support etc. 47 48 * Add a framework to spawn a group of processes in parallel, and use 49 it to run "git fetch --recurse-submodules" in parallel. 50 51 * A slight update to the Makefile to mark "phoney" targets 52 as such correctly. 53 54 * In-core storage of the reverse index for .pack files (which lets 55 you go from a pack offset to an object name) has been streamlined. 56 57 * d95138e6 (setup: set env $GIT_WORK_TREE when work tree is set, like 58 $GIT_DIR, 2015-06-26) attempted to work around a glitch in alias 59 handling by overwriting GIT_WORK_TREE environment variable to 60 affect subprocesses when set_git_work_tree() gets called, which 61 resulted in a rather unpleasant regression to "clone" and "init". 62 Try to address the same issue by always restoring the environment 63 and respawning the real underlying command when handling alias. 64 65 * The low-level code that is used to create symbolic references has 66 been updated to share more code with the code that deals with 67 normal references. 68 69 * strbuf_getline() and friends have been redefined to make it easier 70 to identify which callsite of (new) strbuf_getline_lf() should 71 allow and silently ignore carriage-return at the end of the line to 72 help users on DOSsy systems. 73 74 * "git shortlog" used to accumulate various pieces of information 75 regardless of what was asked to be shown in the final output. It 76 has been optimized by noticing what need not to be collected 77 (e.g. there is no need to collect the log messages when showing 78 only the number of changes). 79 80 81Also contains various documentation updates and code clean-ups. 82 83 84Fixes since v2.7 85---------------- 86 87Unless otherwise noted, all the fixes since v2.7 in the maintenance 88track are contained in this release (see the maintenance releases' 89notes for details). 90 91 * An earlier change in 2.5.x-era broke users' hooks and aliases by 92 exporting GIT_WORK_TREE to point at the root of the working tree, 93 interfering when they tried to use a different working tree without 94 setting GIT_WORK_TREE environment themselves. 95 (merge df1e6ea nd/stop-setenv-work-tree later to maint). 96 97 * The "exclude_list" structure has the usual "alloc, nr" pair of 98 fields to be used by ALLOC_GROW(), but clear_exclude_list() forgot 99 to reset 'alloc' to 0 when it cleared 'nr' to discard the managed 100 array. 101 (merge 2653a8c nd/dir-exclude-cleanup later to maint). 102 103 * Paths that have been told the index about with "add -N" are not 104 quite yet in the index, but a few commands behaved as if they 105 already are in a harmful way. 106 (merge 4d55200 nd/ita-cleanup later to maint). 107 108 * "git send-email" was confused by escaped quotes stored in the alias 109 files saved by "mutt", which has been corrected. 110 (merge 2c510f2 ew/send-email-mutt-alias-fix later to maint). 111 112 * A few unportable C construct have been spotted by clang compiler 113 and have been fixed. 114 (merge a0df2e5 jk/clang-pedantic later to maint). 115 116 * The documentation has been updated to hint the connection between 117 the '--signoff' option and DCO. 118 (merge b2c150d dw/signoff-doc later to maint). 119 120 * "git reflog" incorrectly assumed that all objects that used to be 121 at the tip of a ref must be commits, which caused it to segfault. 122 (merge aecad37 dk/reflog-walk-with-non-commit later to maint). 123 124 * The ignore mechanism saw a few regressions around untracked file 125 listing and sparse checkout selection areas in 2.7.0; the change 126 that is responsible for the regression has been reverted. 127 (merge 8c72236 nd/exclusion-regression-fix later to maint). 128 129 * Some codepaths used fopen(3) when opening a fixed path in $GIT_DIR 130 (e.g. COMMIT_EDITMSG) that is meant to be left after the command is 131 done. This however did not work well if the repository is set to 132 be shared with core.sharedRepository and the umask of the previous 133 user is tighter. They have been made to work better by calling 134 unlink(2) and retrying after fopen(3) fails with EPERM. 135 (merge ea56518 js/fopen-harder later to maint). 136 137 * Asking gitweb for a nonexistent commit left a warning in the server 138 log. 139 140 Somebody may want to follow this up with an additional test, perhaps? 141 IIRC, we do test that no Perl warnings are given to the server log, 142 so this should have been caught if our test coverage were good. 143 (merge a9eb90a ho/gitweb-squelch-undef-warning later to maint). 144 145 * "git rebase", unlike all other callers of "gc --auto", did not 146 ignore the exit code from "gc --auto". 147 (merge 8c24f5b jk/ok-to-fail-gc-auto-in-rebase later to maint). 148 149 * Many codepaths that run "gc --auto" before exiting kept packfiles 150 mapped and left the file descriptors to them open, which was not 151 friendly to systems that cannot remove files that are open. They 152 now close the packs before doing so. 153 (merge d562102 js/close-packs-before-gc later to maint). 154 155 * A recent optimization to filter-branch in v2.7.0 introduced a 156 regression when --prune-empty filter is used, which has been 157 corrected. 158 (merge 1dc413e jk/filter-branch-no-index later to maint). 159 160 * The description for SANITY prerequisite the test suite uses has 161 been clarified both in the comment and in the implementation. 162 (merge 719c3da jk/sanity later to maint). 163 164 * "git tag" started listing a tag "foo" as "tags/foo" when a branch 165 named "foo" exists in the same repository; remove this unnecessary 166 disambiguation, which is a regression introduced in v2.7.0. 167 (merge 0571979 jk/list-tag-2.7-regression later to maint). 168 169 * The way "git svn" uses auth parameter was broken by Subversion 170 1.9.0 and later. 171 (merge 0b66415 ew/svn-1.9.0-auth later to maint). 172 173 * Other minor clean-ups and documentation updates 174 (merge 99487cf ss/user-manual later to maint). 175 (merge e914ef0 ew/for-each-ref-doc later to maint). 176 (merge 36fc7d8 sg/t6050-failing-editor-test-fix later to maint). 177 (merge 60253a6 ss/clone-depth-single-doc later to maint).